Metal β-diketoiminate precursor use in aerosol assisted chemical vapour deposition of gallium- and aluminium-doped zinc oxide

Article abstract of DOI:10.1016/j.poly.2017.10.036

Aerosol assisted chemical vapour deposition (AACVD) has been used to deposit thin films of ZnO from the single-source precursor [Zn(OC(Me)CHC(Me)N(ⁱPr))₂] (1) affording highly transparent (>80%) and conductive films (sheet resistance ～70 KΩ/sq). Extension of this AACVD method whereby related precursors of the type, [R₂M(OC(Me)CHC(Me)N(ⁱPr))] (R = Et, M = Al (2); R = Me, M = Ga (3)), isolated as oils, were added to the precursor solution allowed for the deposition of aluminium- and gallium-doped ZnO (AZO and GZO) films, respectively. Complexes 1–3 were characterised by elemental analysis, NMR and mass spectrometry. Films were deposited in under 30 min at 400 °C, from CH₂Cl₂/toluene solutions with a N₂ carrier gas. Herein we report the bulk resistivity, ρ, of AZO (0.252 Ω cm) and GZO (0.756 Ω cm) films deposited from this novel approach. All the films transparency exceeded 80% in the visible, X-ray diffraction (XRD) showed all films to crystallise in the wurtzite phase whilst X-ray photoemission spectroscopy (XPS) confirmed the presence of the Al and Ga dopants in the films, and highlighted the low C-contamination (<5%) this route offers. Investigation of a mechanism analogous to the Kirkendall effect confirmed that heating of GZO films at 1000 °C produced the spinel structure GaZn₂O₄.

Full text of DOI:10.1016/j.poly.2017.10.036

Polyhedron 140 (2018) 35–41
Contents lists available at ScienceDirect
Polyhedron
journal homepage: www.elsevier.com/locate/poly
Metal b-diketoiminate precursor use in aerosol assisted chemical vapour
deposition of gallium- and aluminium-doped zinc oxide
a
a
b
a
Caroline E. Knapp a, , Caragh Dyer , Nicholas P. Chadwick , Rachael Hazael , Claire J. Carmalt
⇑
a Department of Chemistry, University College London, 20 Gordon Street, Christopher Ingold Laboratories, London WC1H 0AJ, UK
b Centre for Defence Engineering, Cranfield Defence and Security, Cranfield University, Shrivenham, Swindon SN6 8LA, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 18 September 2017
Accepted 27 October 2017
Aerosol assisted chemical vapour deposition (AACVD) has been used to deposit thin films of ZnO from the
single-source precursor [Zn(OC(Me)CHC(Me)N(iPr))2] (1) affording highly transparent (>80%) and con-
ductive films (sheet resistance ꢀ70 K
X/sq). Extension of this AACVD method whereby related precursors
of the type, [R2M(OC(Me)CHC(Me)N(iPr))] (R = Et, M = Al (2); R = Me, M = Ga (3)), isolated as oils, were
added to the precursor solution allowed for the deposition of aluminium- and gallium-doped ZnO
(AZO and GZO) films, respectively. Complexes 1–3 were characterised by elemental analysis, NMR and
mass spectrometry. Films were deposited in under 30 min at 400 °C, from CH2Cl2/toluene solutions with
Keywords:
Zinc oxide
Thin films
Al doping
Ga doping
AACVD
a N2 carrier gas. Herein we report the bulk resistivity, q, of AZO (0.252 X cm) and GZO (0.756 X cm) films
deposited from this novel approach. All the films transparency exceeded 80% in the visible, X-ray diffrac-
tion (XRD) showed all films to crystallise in the wurtzite phase whilst X-ray photoemission spectroscopy
(XPS) confirmed the presence of the Al and Ga dopants in the films, and highlighted the low C-contam-
ination (<5%) this route offers. Investigation of a mechanism analogous to the Kirkendall effect confirmed
that heating of GZO films at 1000 °C produced the spinel structure GaZn2O4.
Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction
Indium tin oxide (ITO) films are the industrial standard with
resistivities of ꢀ1 Â 10À4
X cm, a large bandgap of 4 eV and trans-
Metal oxide thin films have garnered a large amount of atten-
tion in the scientific community due to their wide range of existing
and potential applications, particularly their use in transparent
conducting oxide (TCO) coatings [1]. TCOs are optically transparent
and electrically conductive making them the ideal candidates for
optoelectronic and photovoltaic applications. Currently, the fastest
growing use of TCOs is in flat panel display technology, an area that
covers computer and smartphone screens for consumer electronics
but also for the military and medical industries [2]. There is also
great demand for their use in the glass industry particularly in
electrochromic windows. TCOs often comprise of n-doped tin, zinc
or indium oxides. As a general rule, a TCO should possess a trans-
mittance of >80%, a carrier concentration of 10–20 cmÀ3 and the
host matrix must comprise of a wide band gap metal oxide (>3
eV) and be doped until rendered degenerate to exhibit metallic-
like electrical conductivity [3]. However, these desirable properties
can conflict as the large band gap required for optical transparency
parency in the majority of the visible region. Indeed, the optical
and electrical properties are difficult to match, however alterna-
tives to ITO are widely sought with a view to lowering the cost
and simplifying deposition techniques.
A range of doped zinc oxides are a viable alternative to ITO [4].
The large bandgap, ease of doping and high conductivity of zinc
oxide is consistent with the prerequisites for TCOs however doping
is necessary due to the low carrier concentration and thus high
resistivity of pure ZnO [5]. The doping of zinc oxide occurs through
n-type doping where impurities are introduced into the crystal lat-
tice to cause an excess of negative charge carriers. Doped zinc oxi-
des have been widely explored with dopants, such as arsenic [6],
indium [7] and manganese [8]. Arguably, the most notable dopant
is aluminium, forming Al-doped ZnO (AZO) which is currently used
in the photovoltaic industry, with reported resistivities as low as
2.4 Â 10À4
X cm and a large band gap of 3.4–3.9 eV [9]. Further-
more, although the properties of ITO are considered superior, the
low cost, availability and low toxicity of zinc and aluminium are
appealing industrially. It was reported that AZO exhibits a high sta-
bility at temperatures >700 K and are the only films stable in a gas-
eous plasma containing hydrogen [10] also more recently it has
been reported that combining both aluminium and gallium as
is detrimental to the low resistivity required (ꢀ10À4
X cm), often
causing problems for the doping process.
⇑
Corresponding author.
E-mail address: caroline.knapp@ucl.ac.uk (C.E. Knapp).
https://doi.org/10.1016/j.poly.2017.10.036
0277-5387/Ó 2017 Elsevier Ltd. All rights reserved.

36
C.E. Knapp et al. / Polyhedron 140 (2018) 35–41
dopants offers a balance between chemical stability and conductiv-
ity [11]. Various doping concentrations have been explored in the
literature with most lying between 2 and 10at.%. Muiva et al.
reported the optimum doping concentration to be 2 at.% for
enhanced conductivities and transmittance >85% [12].
Previous reports of AZO and GZO synthesis describe chlorine
contamination as a result of precursor ligand design, in this paper,
to circumnavigate this we combine the use of an identical b-dike-
toiminate ligand motif on zinc, aluminium and gallium compounds
with AACVD in order to facilitate doping into the ZnO lattice. The
use of the solution based technique, AACVD, overcomes the need
for volatile precursors whilst the metal b-diketoiminate precursors
presented herein exhibit similar decomposition mechanisms. The
precursors therefore decompose cleanly, readily dope the ZnO
framework (at 400 °C) and produce increasingly conductive thin
films of ZnO, GZO and AZO, respectively, free from chlorine
contamination.
Group III elements are common dopants for ZnO [13] and
moving down the group from Al to Ga as
a dopant gives
Ga-doped zinc oxide (GZO), another alternative to ITO. Studies
suggest that gallium doping is more effective than aluminium
doping as the lower reactivity of gallium allows it to be easily
controlled during the doping process. Additionally, the smaller
Ga-O bond length minimises the deformity of the ZnO crystal
lattice allowing higher concentrations of the dopant to be intro-
duced into the host matrix. Furthermore, with a reported resis-
2. Material and methods
tivity between
4
and 5 Â 10À4
X cm it also offers similarly
impressive electrical properties [14,15]. In addition to dopant
concentration, electrical properties of films can also depend on
film thickness. Fortunato et al. discovered that increasing the
GZO film thickness also increased the electron mobility as
defects are more common in thinner films thus scattering charge
carriers. However, saturation of the resistivity occurred at thick-
nesses that exceeded 500 nm [15].
2.1. Precursor synthesis
ZnEt2, AlEt3 and GaMe3 are pyrophoric substances that can
ignite spontaneously in air and react violently with water. There-
fore, all reactions involving air sensitive materials were carried
out under a dry dinitrogen (99.99% from BOC) atmosphere using
an MBraun glove box and Schlenk techniques. All experimental
should be conducted in a fume hood. Following the deposition
films were air and moisture stable and were safe to handle as
any reactive species leave via the reactor exhaust during the
AACVD process. All solvents used were stored in alumina columns
and dried, such that the water concentration was 5–10 ppm. AlEt3
and GaMe3 were procured from SAFC Hitech, ZnEt2 and all other
chemicals were purchased from Aldrich and stored appropriately.
1H and 13C NMR spectra were obtained on a Bruker AMX500
spectrometer, operating at 500.13 MHz, using CD2Cl2 or CDCl3
which were dried and degassed over molecular sieves prior to
use; 1H and 13C chemical shifts are reported relative to SiMe4 (d
0.00). IR spectra were recorded using a Shimadzu FTIR-8200 spec-
trometer, operating in the region of 4000–400 cmÀ1. Mass spectra
were obtained using a Micromass 70-SE spectrometer using Chem-
ical Ionoisation (CI) with methane reagent gas. Elemental Analysis
was carried using Elemental Analyser (CE-440) (Exeter Analytical
Inc). The instrument used for thermal analysis was a Netzsch Jupi-
ter. All measurements were carried out with the precursor sample
sealed in an aluminium pan. The data was recorded from room
temperature to 600 °C (see S.I.).
Recently studies of films of AZO and GZO deposited via AACVD
showed the correlation between dopant levels and resultant resis-
tivity [16,17]. Commercially procured zinc acetylacetonate, alu-
minium chloride and gallium chloride were used to synthesize
ZnO doped with 5, 10, 15 and 20 at.% Al or Ga. The resultant films
showed transparency greater than 80% and resistivities in the
order of 10À3
X cm [18]. Earlier this year both Al (1.5 at.%) and
Ga (1.5 at.%) doped ZnO powders and thin films (produced via
AACVD) were also reported with resistivity values of 5.6 Â 10À4
X
cm and 5.7 Â 10À3
X cm, respectively using a microwave
assisted synthesis [19]. In both of these reports, where commer-
cially available precursors were used, carbon and/or chlorine con-
tamination was reported in the resultant materials.
Research into metal oxide thin film synthesis has been carried
out hand in hand with the development of metal–organic precur-
sors, usually metal alkoxides or b-diketonates designed to limit
contamination [20]. These are single-source precursors containing
a direct metal–oxygen bond that can either be synthesised or in
some cases are commercially available. b-ketoimine ligands can
form a complex with a metal centre, forming a delocalised ring
including O and N donor atoms. The use of the ligand has been
reported to increase thermal stability and lower the melting point.
The move from b-diketonates to b-diketoiminate facilitates the
incorporation of a N atom into the delocalised ring which can be
further functionalised – allowing various properties of the precur-
sors to be tuned with the careful selection of R group on the N [21].
It has been shown recently that varying the R group on the nitro-
gen atom in a range of b-diketoiminate ligands dramatically
altered the properties of the resultant aluminium and gallium con-
taining precursors [22].
The ability to isolate a precursor as an oil, can be advantageous,
particularly for use in AACVD, where precursors must be soluble in
order to generate the aerosol mist [9]. Ligand modification is not
limited to functionalisation on the nitrogen atom, it can include
the fluorination of the alkyl backbone; these electron withdrawing
groups can reduce the strength of the intermolecular forces via
increased electron density at the metal centre resulting in a more
volatile product. However contamination and the risk of metal cor-
rosion from fluoride species must be taken into consideration
[3,20]. A range of zinc b-diketoiminate compounds have been suc-
cessfully utilized in the CVD of ZnO films [23,24], and it is proposed
that an extension to this groundwork could include the incorpora-
tion of Al and Ga dopants to improve the conductivity of the resul-
tant materials.
Synthesis of the protonated ligand: [(OC(Me)CHC(Me)NH(iPr))],
[Zn(OC(Me)CHC(Me)N(iPr))2] (1) and [Et2Al(OC(Me)CHC(Me)N
(iPr))] (2) are consistent with literature and are included in S.3.
2.2. Synthesis of [Me2Ga(OC(Me)CHC(Me)N(iPr))] (3)
The b-ketoiminate ligand, [(OC(Me)CHC(Me)NH(iPr))] (1.00 g,
7.0 mmol) in toluene (20 mL) was added dropwise to a solution
of GaMe3 (0.81 g, 7.0 mmol) in toluene (20 mL) at À78 °C. This
was stirred overnight and the solvent was removed in vacuo to
yield a viscous yellow oily product 1 (70%). 1H NMR d/ppm
(C6D6): 4.56 (s, 1H, COCH), 3.21 (m, 1H, CH(Me)2), 1.66 (s, 3H,
MeCO), 1.34 (s, 3H, CNMe), 1.10 (6H, d, NCH(Me)2), 0.14 (4H, m,
GaMe). 13C{1H} NMR d/ppm 180.0 (CO), 165.2 (CN), 82.3 (CH),
39.7 (NC(Me)2), 26.2 (OCMe), 23.5 (NC(Me)2), 22.2 (COMe), 1.8
(GaMe). M/S: m/z [ES] + 239.99. Elemental Anal. calc. %: C: 50.05,
H: 8.40, N: 5.84. Found: C: 51.3, H: 9.1, N: 5.99%.
2.3. Chemical vapour deposition
AACVD reactions took place within a fume hood. Depositions
were carried out on SiO2 coated barrier glass substrates with pro-
portions: 90 mm  45 mm  4 mm within a cold walled, horizon-

C.E. Knapp et al. / Polyhedron 140 (2018) 35–41
37
tal bed reactor. Two glass substrates were used to form a bottom
plate which sat directly on top of the graphite heating block and
a top plate positioned above it with a 0.5 cm gap between them.
The desired thin film was deposited on the bottom plate, whilst
the top plate reduces turbulence, ensuring laminar gas flow. The
pounds 1 and 3 (also one-pot) in a ratio of 1:0.05 (0.5 mmol scale).
The experimental procedure described for GZO was repeated with
the addition of a piece of glass quartz (2 cm  2 cm) placed on top
of the substrate. The quartz was then transferred to a Carbolite cwf
13/13 furnace and annealed at 1000 °C overnight, in order to carry
out XRD analysis.
i
glass substrates were cleaned thoroughly using water, PrOH and
acetone and then dried in air before each use. The temperature
of the graphite heating block was controlled using a platinum–rho-
dium thermocouple and the reactor was kept under a flow of the
N2 carrier gas and the aerosol entering through the brass baffle.
The reactor was sealed with a cylindrical quartz tube capped by
two stainless steel end plates with an exhaust. Before the aerosol
was formed, the precursor was weighed out into a glass bubbler
and appropriate solvent added. Once the reactor had reached the
required temperature, the glass bubbler was attached to the reac-
tor using PTFE tubing via a one way glass T-piece. The flow meter,
controlling the N2 carrier gas flow, was diverted through the T-
piece diverting the aerosol mist generated by a humidifier into
the reactor via the baffle. To create the aerosol, an ultrasonic
humidifier containing a piezoelectric device, which functions at
20 kHz was used. XRD was carried out with a Bruker D8 Discover
X-ray diffractometer using monochromatic Cu K 1 and Cu K 2 radi-
3. Results and discussion
3.1. Synthesis of 1–3
b-Ketoimine compounds, formed in a simple condensation reac-
tion have been widely explored as ligands for a range of metals and
used in the synthesis of successful precursors for metal oxide thin
films [22,23]. Following the synthesis and isolation of the ligand,
[(Me)CN(H){iPr}–CHC(Me) = O], two equivalents were reacted with
diethyl zinc in toluene and a white precipitate of 1 was obtained
and isolated in good yield, after being filtered and washed with
hexane. Conversely compounds 2 and 3 were not isolated as pre-
cipitates after an overnight reflux of a 1:1 reaction of the afore-
mentioned ligand with AlEt3 and GaMe3 which yielded oils of 2
and 3, respectively. Formation of 1 and 2 were confirmed via 1H,
13C, MS and EA and were consistent with literature [22,23]. The
proton resonances in the NMR of 3 for the Ga-ligand environments
were significantly shifted from 1H environments observed for the
protonated ligand: [(Me)CN(H){iPr}–CHC(Me) = O] alone, evidenc-
ing the formation of 3. Elemental analysis and mass spectrometry
confirm the isolation of 3, furthermore, the 2:1 ratio of methyl
group proton environments to the Ga-ligand proton environments
in the 1H NMR support the formation of 3 (Scheme 1).
a
a
ation of wavelengths 1.54056 and 1.54439 Å respectively, emitted
in an intensity ratio of 2:1, voltage = 40 kV; current = 40 mA. SEM
was performed using a Philips XL30 FEG with an electron beam
accelerating energy of 30 kV. XPS profiling was performed using
a Thermo Scientific K-Alpha XPS system using monochromatic Al
K
radiation at 1486.6 eV X-ray source. CasaXPS software was used
a
to analyse the data with binding energies referenced to an adven-
titious C 1s peak at 284.8 eV. UV/Vis/NIR transmission spectra
were recorded using a PerkinElmer Lambda 950 spectrometer in
the range of 300–1400 nm with an air background.
2.4. General procedures
3.2. AACVD of ZnO, AZO and GZO
Due to the air sensitive nature of the precursors, standard Sch-
lenk and glovebox (MBraun Unilab) techniques were employed
and all precursors and solvents were kept under a dry N2 atmo-
sphere. Solvents were added to the vessel containing the precursor
using an oven-dried cannula and then attached to the AACVD
apparatus under a flow of N2. A number of different flow rates were
tested ranging between 1 and 2 L/min as well as a variety of differ-
ent deposition temperatures. The AACVD reaction lasted between
30 and 60 min and the reactor was set to cool to room tempera-
ture, still under a flow of N2. Many attempts were made to opti-
mise the AACVD reaction of the precursor testing different
solvents, flow rates, temperatures and masses of precursor.
Following optimisation of the AACVD process (see S.I.) com-
pound 1 was used to deposit thin films of ZnO at 400 °C. Films of
AZO were achieved at 400 °C via a one-pot AACVD charged with
1 and 2 in a 1:0.05 ratio. Following the procedure for the synthesis
of thin films of AZO (at 400 °C), GZO films were also successfully
deposited using compounds 1 and 3 in a ratio of 1:0.05. For all
films the CH2Cl2/toluene solvent system was utilised (see S.I.).
The films of ZnO, AZO and GZO were transparent displaying a slight
brown tinge, indicative of carbon contamination. All films were
adherent to the glass substrate after wiping, scratching with a scal-
pel, the scotch tape test and treating with acetone and other
organic solvents. The films perished when left in nitric acid
overnight.
2.5. AACVD of ZnO from 1
The XRD pattern for films deposited from 1 show alignment to
peaks typical of zinc oxide (Fig. 1), with preferential orientation for
the (002) peak indicating orientation along the c axis [25]. Using
the Le Bail method in GSAS, the lattice parameters of the ZnO films
were determined to be a = 3.2367(5) Å and b = 5.1613(5) Å.
AZO films deposited from a one-pot mixture of 1 and 2 show a
significant decrease in peak intensity with Al doping in their XRD
pattern (Fig. 1). This may be due to the varying film thickness
which is well known to affect the intensity of the XRD peaks. A
shift in the peak position when doping occurs due to the change
in lattice parameters that accompanies the introduction of dopant
atom into the host matrix, shifting was observed when compared
to the 2h values of the undoped zinc oxide films. Firstly for the
(002) peak from 34.85° to 35.00° 2h and the (100) peak from
31.40° to 32.15° 2h. The replacement of Zn2+ ions with the smaller
Al3+ ions caused a shift of the (100) peak to a larger 2h [26]. Lattice
parameters of the unit cell were calculated for AZO films to inves-
tigate the effect of doping on the crystal structure. Using the afore-
1 (0.2 g, 0.578 mmol) was added to a glass flask in the glove
box. A small amount of CH2Cl2 (ca. 10 mL) was added to dissolve
the precursor followed by toluene (ca. 30 mL). The resultant solu-
tion was a transparent, pale yellow/orange colour. The AACVD
reactor was heated to 400 °C whilst under a flow of N2 gas at a flow
rate of 1 L/min. The flow of N2 was then diverted into the glass
bubbler through the precursor solution allowing the aerosol mist
to be directed into the AACVD reactor. The deposition took 30
min and the reactor was cooled to room temperature under a flow
of N2 gas before the substrate was removed.
2.6. AACVD of AZO, GZO and GaZn2O4
The experimental procedure described above was repeated sev-
eral times, films of AZO from one-pot of 1 and 2 in a 1:0.05 ratio
(0.5 mmol scale). GZO films were deposited as above using com-

38
C.E. Knapp et al. / Polyhedron 140 (2018) 35–41
Scheme 1. Synthesis of b-ketoiminate metal complexes 1–3 and the conditions for the AACVD of thin films of ZnO, AZO and GZO.
reported in the literature that incorporation into the ZnO crystal
structure causes the (002) to shift to a higher angle side, reflected
in the decrease in the c parameter from 5.16 Å to 5.15 Å [27].
Energy dispersive X-ray (EDX) analysis of films of ZnO depos-
ited from 1 confirmed the presence of zinc but could not be used
to give a metal to oxygen ratio due to breakthrough to the sub-
strate. Carbon contamination was apparent from the EDX data
(ꢀ15 at.% carbon). The EDX analysis of the AZO films confirmed
the Al content of the films to be 4.8 at.%, closely corroborating
the experimental concentrations used. Films of GZO deposited
from 1 and 3 were analysed by EDX and shown to contain 4.5 at.
% Ga, in both the AZO and GZO films carbon content was below
5 at.% (approaching detection limit) no chlorine contamination
could be detected.
In the X-ray photoelectron spectroscopy (XPS) data all peaks
were fitted by a Gaussian/Lorentzian product distribution. For
films deposited from 1 a Zn:O ratio of 1.1:1.0 was calculated,
indicative of binary ZnO with a Zn 2p3/2 binding energy of
1021.5 eV [28]. In Fig. 2 (top left) the Zn 2p peaks are offset for
Fig. 1. XRD pattern showing: ZnO deposited by AACVD of 1 at 400 °C (black); [10]
AZO deposited from 1 and 2 (red); GZO deposited from 1 and 3 (blue). Peaks verify
the hexagonal wurtzite structure expected. (Colour online.)
ZnO (black), AZO (red) and GZO (blue), a slight shift, consistent
with the presence of Al and Ga dopants, respectively, was observed
for the Zn 2p3/2 binding energy: 1022.7 eV (AZO) and 1022.3 eV
(GZO). The O 1s spectra recorded for the ZnO, AZO and GZO films
(Fig. 2, top right) all included a peak located between 530 and
534 eV. This can be assigned to O2- ions in the ZnO hexagonal wurt-
zite structure. The appearance of a shoulder in each of the spectra
correspond to the presence of loosely bound surface species.
XPS confirmed the presence of Al(III) in the AZO films deposited
from 1 and 2, the binding energy of Al 2p3/2 was 75.1 eV and the Al
2p1/2 was 75.5 eV (Fig. 2, bottom left), higher than previous reports
for Al2O3 [22], which is consistent for Al as a dopant as opposed to
when incorporated within its binary oxide [29]. In the XPS of the
GZO films deposited from 1 and 3, the binding energy of the Ga
3d5/2 at 20.2 eV is indicative of Ga(III) (Fig. 2, bottom right). Since
this technique only probes the surface and not the bulk of the
material and this, combined with the issues associated with the
low doublet splitting in the Al 2p and Ga 3d ionizations, at.% of alu-
minium or gallium could not be calculated from the XPS data [30].
Using UV/Vis spectroscopy (Fig. 3), it was possible to obtain
information using transmission reflectance data about the trans-
parency and bandgap of the thin films of ZnO, AZO and GZO by
using a Tauc plot. A value of 3.13 eV was calculated in accordance
with literature values for the band gap of the thin films of ZnO
deposited using 1 [9]. Comparatively the band gap of the AZO films
deposited from compounds 1 and 2 and the GZO films deposited
mentioned methods the lattice parameters were determined to be
a = 3.2353(5) Å and b = 5.1785(5) Å.
In the XRD patterns of films of GZO deposited from one-pot
mixtures of 1 and 3 preferential orientation remained for the
(002) peak, and the intensity of the peaks was decreased in com-
parison with undoped ZnO (Fig. 1). A very slight shift in the
(002) peak was observed in the spectrum from 34.85° in the
undoped ZnO film to 34.9° 2h in the GZO. However, the difference
in size between the ionic radii of zinc and gallium is relatively
small, therefore shifting may not occur as the distortion of the
crystal lattice could be minimal, with similar observations have
been reported in the literature [14]. The shift is smaller than that
observed for the AZO films which is expected due to the smaller
difference in ionic radii between zinc and gallium compared to that
of zinc and aluminium.
As with AZO, lattice parameters of the unit cell were calculated
for GZO films to investigate the effect of doping on the crystal
structure. Using the Le Bail method lattice parameters were deter-
mined to be a = 3.2257(5) Å and b = 5.1496(5) Å. The data reveals a
slight shift in the parameters showing a small effect on the crystal
structure by the dopant atoms. In the case of gallium, it is well

C.E. Knapp et al. / Polyhedron 140 (2018) 35–41
39
Fig. 2. XPS analysis of: Zn 2p peaks in ZnO (black), AZO (red) and GZO (blue) (top left); O 1s peaks in ZnO (black), AZO (red) and GZO (blue) (top right); Al 2p in AZO (red)
(bottom left); and Ga 3d in GZO (blue) (bottom right). (Colour online.)
from 1 and 3 were found to be 3.25 and 3.28 eV respectively. The
increase in the bandgap upon doping may be expected due to an
increased charge carrier concentration due to the Burnstein-Moss
effect widening the bandgap [14]. The slightly wider band gap con-
tributes to the high optical transparency of the films. The data
showed transmission exceeded 80% in all films in the visible region
fitting the requirements for a TCO.
Resistance of the films was determined by two-point probe,
additionally, using the film thickness obtained from Filmetrics
equipment the resistivity was calculated. Table 1 summarises the
data for the ZnO, AZO and GZO films [26].
trical properties significantly (Table 1). However, the effect of film
thickness must be noted and although not a direct correlation,
thicker films generally had better conductivity.
Scanning electron microscopy (SEM) shows the surface mor-
phology of the zinc oxide films deposited from 1 (Fig. 4, black bor-
der) to be similar in appearance from the AZO films indicating that
doping, in this instance seems to have little effect on the morphol-
ogy. Interestingly however the GZO films appeared to be formed of
larger spherical units (Fig. 4, blue border). These GZO films were
also deposited on quartz and then annealed at 1000 °C. The rela-
tively equally sized clusters (ꢀ70 nm) observed on the GZO films
were replaced with an array of crystalline blocks, after annealing
(Fig. 4, green border).
A lower resistivity was recorded in the doped films compared to
the ZnO indicating that the addition of dopant improved the elec-
Table 1
Summary of the characterisation of ZnO, AZO and GZO thin films: film thickness (d), sheet resistance (RSh), bulk resistivity (q), cell parameters were calculated using a Le Bail
refinement, position of the (002) peak in the XRD pattern (°), band gap (Eg), at.% of dopant from EDX.
Film
ZnO
d (nm)
182.6
RSh (k
X
sqÀ1
)
q
(
X
cm)
Cell parameters (Å)
(002) Peak position (°)
Eg (eV)
3.13
EDX (at.%)
–
70
1.270
0.252
0.756
a = 3.2367(5)
b = 5.1613(5)
a = 3.2353(5)
b = 5.1785(5)
a = 3.2257(5)
b = 5.1496(5)
34.85
AZO
GZO
210.3
189.1
12
40
35.00
34.90
3.25
3.28
4.8 (Al)
4.5 (Ga)

40
C.E. Knapp et al. / Polyhedron 140 (2018) 35–41
separation in the XPS for gallium implies that any number of envi-
ronments can fit the single peak recorded.
In order to investigate the effect this layer may have on the
material beneath, films of GZO were annealed at 1000 °C. The
XRD data below (Fig. 5) confirms the spinel zinc gallate, ZnGa2O4
formed alongside Ga2O3. Annealing the film had therefore caused
the formation of ZnGa2O4 from a reaction between Ga2O3 present
on the surface and the underlying GZO [29].
Ga2O3 and GaZn2O4 are the thermodynamically stable products
at 1000 °C when compared to GZO. These results confirm AACVD as
a simple alternative to some multistep processes [5] which are cur-
Fig. 3. UV/vis data for ZnO thin film deposited from 1 at 400 °C (black); AZO film
deposited from 1 and 2 at 400 °C (red); GZO film deposited from 1 and 3 at 400 °C
(blue). (Colour online.)
3.3. Investigation of the Kirkendall effect
The Kirkendall effect – namely bulk diffusion at the interface of
two metal layers which produces both vacancies and new materi-
als – has been exploited in recent years for the synthesis of hollow
nanomaterials [31,32]. A mechanism analogous to this effect has
been used to synthesize spinel structures of the type AB2O4, from
a range of techniques including ALD and CVD whereby layers of
different materials are heated above a certain temperature [33–
35]. It is likely that the GZO films presented herein have an amor-
phous layer of oxide at their surface. This can be difficult to detect
in films deposited at 400 °C because Ga2O3 is amorphous at this
temperature ruling out XRD analysis, furthermore the low doublet
Fig. 5. XRD spectra of ZnO on glass (black); GZO film deposited on glass (blue) and
GZO on quartz annealed at 1000 °C (green). Diamond markers in the annealed
spectrum show where Bragg peaks occur characteristic of zinc gallate, triangle
markers show Ga2O3. (Colour online.)
Fig. 4. SEM images of thin films grown at 400 °C of: ZnO deposited from 1 (black border); b) AZO deposited from a one-pot mixture of 1 and 2 (red border); GZO deposited
from a one-pot mixture of 1 and 3 (blue border); and the crystalline spinel: GaZn2O4 produced from heating GZO thin films on quartz substrates at 1000 °C (green border).
(Colour online.)

C.E. Knapp et al. / Polyhedron 140 (2018) 35–41
41
[13] L. Schmidt-Mende, J.L. MacManus-Driscoll, ZnO – nanostructures, defects, and
devices, Mater. Today 10 (2007) 40, https://doi.org/10.1016/S1369-7021(07)
70078-0.
rently used to deposit spinels consistent with previous investiga-
tions into post-annealing temperatures [36].
[14] M.-C. Jun, S.-U. Park, J.-H. Koh, Comparative studies of Al-doped ZnO and Ga-
doped ZnO transparent conducting oxide thin films, Nanoscale Res. Lett. 7
(2012) 639, https://doi.org/10.1186/1556-276X-7-639.
4. Conclusions
[15] E. Fortunato, D. Ginley, H. Hosono, D.C. Paine, Transparent conducting oxides
for photovoltaics, MRS Bull. 32 (2007) 242, https://doi.org/10.1557/
mrs2007.29.
[16] S. Chen, G. Carraro, D. Barreca, R. Binions, Growth and electro-optical
properties of Ga-doped ZnO films prepared by aerosol assisted chemical
vapour deposition, Thin Solid Films 584 (2015) 316, https://doi.org/10.1016/j.
tsf.2014.11.092.
Following the synthesis of zinc, aluminium and gallium con-
taining compounds using the same b-ketoimine ligand – each con-
taining pre-formed metal–oxygen bonds, AACVD has been used to
deposit in a simple one step procedure thin films of ZnO, AZO and
GZO. Films were deposited on glass at 400 °C, and in the case of
GZO, also on quartz for post-annealing treatment confirming the
formation of the thermodynamically favourable spinel GaZn2O4.
Dopant concentration in the films matches the amount added to
the reaction pot, proving this to be an efficient and non-wasteful
technique. In conclusion the thin films deposited using these pre-
cursors showed reasonable electrical conductivity and showed
excellent optical properties; a promising proof of concept for
one-pot made-to-measure precursor systems.
[17] S.D. Ponja, S. Sathasivam, I.P. Parkin, C.J. Carmalt, Transparent conductive
aluminium and fluorine co-doped zinc oxide films via aerosol assisted
chemical vapour deposition, RSC Adv.
10.1039/C4RA09997D.
4 (2014) 49723, https://doi.org/
[18] D.B. Potter, D.S. Bhachu, M.J. Powell, J.A. Darr, I.P. Parkin, C.J. Carmalt, Al-, Ga-,
and In-doped ZnO thin films via aerosol assisted CVD for use as transparent
conducting oxides: Al-, Ga-, and In-doped ZnO thin films via aerosol assisted
CVD, Phys. Status Solidi
pssa.201532996.
A 213 (2016) 1346, https://doi.org/10.1002/
[19] D.S.Y. Jayathilake, T.A.N. Peiris, J.S. Sagu, D.B. Potter, K.G.U. Wijayantha, C.J.
Carmalt, D.J. Southee, Microwave-assisted synthesis and processing of Al-
doped, Ga-doped, and Al, Ga codoped ZnO for the pursuit of optimal
conductivity for transparent conducting film fabrication, ACS Sustainable
Chem. Eng. 5 (2017) 4820, https://doi.org/10.1021/acssuschemeng.7b00263.
[20] D. Pugh, P. Marchand, I.P. Parkin, C.J. Carmalt, Group 13 b-ketoiminate
compounds: gallium hydride derivatives as molecular precursors to thin films
of Ga2O3, Inorg. Chem. 51 (2012) 6385, https://doi.org/10.1021/ic3006794.
[21] P. Marchand, C.J. Carmalt, Molecular precursor approach to metal oxide and
pnictide thin films, Coord. Chem. Rev. 257 (2013) 3202, https://doi.org/
10.1016/j.ccr.2013.01.030.
Acknowledgements
C.E.K. thanks the Ramsay Memorial Trust for funding. C.J.C
thanks the EPSRC for the grant: EP/K001515.
[22] C.E. Knapp, P. Marchand, C. Dyer, I.P. Parkin, C.J. Carmalt, Synthesis and
characterisation of novel aluminium and gallium precursors for chemical
vapour deposition, New J. Chem. 39 (2015) 6585, https://doi.org/10.1039/
C5NJ00947B.
[23] J.S. Matthews, O.O. Onakoya, T.S. Ouattara, R.J. Butcher, Synthesis and
characterization of zinc AP-MOCVD precursors and their utility in the
growth of ZnO, Dalton Trans. (2006) 3806, https://doi.org/10.1039/b603308c.
[24] J. Holmes, K. Johnson, B. Zhang, H.E. Katz, J.S. Matthews, Metal organic
chemical vapor deposition of ZnO from b-ketoiminates: MOCVD of ZnO from
b-ketoiminates, Appl. Organomet. Chem. 26 (2012) 267, https://doi.org/
10.1002/aoc.2850.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at https://doi.org/10.1016/j.poly.2017.10.036.
References
[1] C.E. Knapp, C.J. Carmalt, Solution based CVD of main group materials, Chem.
Soc. Rev. 45 (2016) 1036, https://doi.org/10.1039/C5CS00651A.
[25] Z.R. Khan, Optical and structural properties of ZnO thin films fabricated by sol-
gel method, Mater. Sci. Appl. 02 (2011) 340, https://doi.org/10.4236/
msa.2011.25044.
[2] B.G. Lewis, D.C. Paine, Applications and processing of transparent conducting
oxides, MRS Bull. 25 (2000) 22, https://doi.org/10.1557/mrs2000.147.
[3] Y.-H. Liu, Y.-C. Cheng, Y.-L. Tung, Y. Chi, Y.-L. Chen, C.-S. Liu, S.-M. Peng, G.-H.
Lee, Synthesis and characterization of fluorinated ??-ketoiminate and imino-
alcoholate Pd complexes: precursors for palladium chemical vapor deposition,
J. Mater. Chem. 13 (2003) 135, https://doi.org/10.1039/b208535f.
[4] S.-U. Park, J.-H. Koh, Electrical and optical properties of in and Al doped ZnO
thin film, Electron. Mater. Lett. 9 (2013) 493, https://doi.org/10.1007/s13391-
013-0046-9.
[5] E. Fortunato, L. Raniero, L. Silva, A. Goncalves, A. Pimentel, P. Barquinha, H.
Aguas, L. Pereira, G. Goncalves, I. Ferreira, Highly stable transparent and
conducting gallium-doped zinc oxide thin films for photovoltaic applications,
Sol. Energy Mater. Sol. Cells. 92 (2008) 1605, https://doi.org/10.1016/
j.solmat.2008.07.009.
[26] S.S. Meysami, A.A. Koós, F. Dillon, N. Grobert, Aerosol-assisted chemical vapour
deposition synthesis of multi-wall carbon nanotubes: II. An analytical study,
Carbon 58 (2013) 159, https://doi.org/10.1016/j.carbon.2013.02.041.
[27] S. Liang, Z.-X. Mei, X.-L. Du, Modulation of electrical and optical properties of
gallium-doped ZnO films by radio frequency magnetron sputtering, Chin. Phys.
B 21 (2012) 067306, https://doi.org/10.1088/1674-1056/21/6/067306.
[28] H.A. Ali, A.A. Iliadis, A. Von Cresce, P. Kofinas, U. Lee, Properties of self-
assembled ZnO nanostructures on Si and SiO/sub 2/ wafers, in: IEEE, 2001, 454,
doi:10.1109/ISDRS.2001.984543.
[29] J.-L. Yan, Y.-N. Zhao, C. Li, Formation of ZnGa2O4 films by multilayer deposition
and subsequent thermal annealing, Chin. Phys. B 23 (2014) 048105, https://
doi.org/10.1088/1674-1056/23/4/048105.
[6] S.Y. Li, P. Lin, C.Y. Lee, T.Y. Tseng, C.J. Huang, Effect of Sn dopant on the
properties of ZnO nanowires, J. Phys. Appl. Phys. 37 (2004) 2274, https://doi.
org/10.1088/0022-3727/37/16/009.
[7] A. Hafdallah, F. Yanineb, M.S. Aida, N. Attaf, In doped ZnO thin films, J. Alloys
Compd. 509 (2011) 7267, https://doi.org/10.1016/j.jallcom.2011.04.058.
[8] V.R. Shinde, T.P. Gujar, C.D. Lokhande, R.S. Mane, S.-H. Han, Mn doped and
undoped ZnO films: A comparative structural, optical and electrical properties
study, Mater. Chem. Phys. 96 (2006) 326, https://doi.org/10.1016/
j.matchemphys.2005.07.045.
[30] J. Creighton, P. Ho, Introduction to Chemical Vapor Deposition (CVD), ASM
International, Materials Park, OH, 2001.
[31] H.J. Fan, M. Knez, R. Scholz, D. Hesse, K. Nielsch, M. Zacharias, U. Gösele,
Influence of surface diffusion on the formation of hollow nanostructures
induced by the Kirkendall effect: the basic concept, Nano Lett. 7 (2007) 993,
https://doi.org/10.1021/nl070026p.
[32] H. Jin fan, M. Knez, R. Scholz, K. Nielsch, E. Pippel, D. Hesse, M. Zacharias, U.
Gösele, Monocrystalline spinel nanotube fabrication based on the Kirkendall
effect, Nat. Mater. 5 (2006) 627, https://doi.org/10.1038/nmat1673.
[33] C.E. Knapp, J.A. Manzi, A. Kafizas, I.P. Parkin, C.J. Carmalt, Aerosol-assisted
chemical vapour deposition of transparent zinc gallate films, ChemPlusChem
79 (2014) 1024, https://doi.org/10.1002/cplu.201402037.
[9] K. Ellmer, Resistivity of polycrystalline zinc oxide films: current status and
physical limit, J. Phys. Appl. Phys. 34 (2001) 3097, https://doi.org/10.1088/
0022-3727/34/21/301.
[10] M.N. Islam, T.B. Ghosh, K.L. Chopra, H.N. Acharya, XPS and X-ray diffraction
studies of aluminum-doped zinc oxide transparent conducting films, Thin
Solid Films 280 (1996) 20, https://doi.org/10.1016/0040-6090(95)08239-5.
[11] D.P. Howard, P. Marchand, L. McCafferty, C.J. Carmalt, I.P. Parkin, J.A. Darr,
High-throughput continuous hydrothermal synthesis of transparent
conducting aluminum and gallium co-doped zinc oxides, ACS Comb. Sci. 19
(2017) 239, https://doi.org/10.1021/acscombsci.6b00118.
[34] C.E. Knapp, I.D. Prassides, S. Sathasivam, I.P. Parkin, C.J. Carmalt, Aerosol-
assisted chemical vapour deposition of
a copper gallium oxide spinel,
ChemPlusChem 79 (2014) 122, https://doi.org/10.1002/cplu.201300289.
[35] S. Kumar, G. Sarau, C. Tessarek, M. Göbelt, S. Christiansen, R. Singh, Study of
high quality spinel zinc gallate nanowires grown using CVD and ALD
techniques, Nanotechnology 26 (2015) 335603, https://doi.org/10.1088/
0957-4484/26/33/335603.
[12] C.M. Muiva, T.S. Sathiaraj, K. Maabong, Effect of doping concentration on the
properties of aluminium doped zinc oxide thin films prepared by spray
pyrolysis for transparent electrode applications, Ceram. Int. 37 (2011) 555,
https://doi.org/10.1016/j.ceramint.2010.09.042.
[36] C.H. Ahn, S.H. Kim, Y.K. Kim, H.S. Lee, H.K. Cho, Effect of post-annealing
temperatures on thin-film transistors with ZnO/Al2O3 superlattice channels,
Thin Solid Films 584 (2015) 336, https://doi.org/10.1016/j.tsf.2015.01.017.